Digital Leaders AI Pulse Issue #2

Welcome to the latest edition of The AI Pulse for Digital Leaders. A curated collection of essential articles, commentaries, and news stories from reputable sources that bring insight to digital leaders on the principles and practices of delivering AI-at-Scale.

Highlights from this edition: The benefits of using GenAI and digital twins technology together; A summary of 185 GenAI use cases; and a discussion on Frugal AI.

AI for Good

Adoption of AI at Scale is not just for very large companies and governments. Here, the Forum of Small States (FOSS) has issued a very useful document with guidance, frameworks, and real examples of AI adoption within small states such as Singapore and Rwanda. It contains a very interesting set of ideas for addressing concerns such as managing limited budgets, establishing trust, and creating broader cooperative partnerships.

Stanford’s Digital Economy Lab provides a summary of how a multidisciplinary group of thinkers see the future of AI and governance in volume called The Digitalist Papers. Themes span AI and governance, AI and civic engagement, AI regulation, and AI and democratic values.

An interesting set of comments from Kenneth Payne on why his excitement about OpenAI's progress in developing AI with emotional insight has turned to disappointment. Payne believes that the decision to prohibit AI from expressing emotions decision has significant implications for the future of AI and human society, and that Europe needs to have a more inclusive debate about the potential benefits and risks of this technology.

Bias and Ethics

The International Bar Association (IBA) and the Center for AI and Digital Policy (CAIDP) have launched a new report that explores the transformative impact of AI on the legal profession and offers critical insights into the governance and ethical deployment of AI technologies in legal practice and how they help society.

AI models continue to grow in complexity. This short Forbes article reviews how understanding their decision-making processes—referred to as model interpretability—has become critical for ensuring reliability, accountability and responsible use of AI.

Cyber Security

This year’s Turing Lectures began this week with Jonathan Bright, Head of Online Safety at The Alan Turing Institute, exploring AI's role in political campaigns, deepfakes and content moderation. His lecture focused on AI's impact on democracy and the importance of scepticism when consuming digital media. More lectures on these topics to follow.

Funded by the Carnegie Endowment for International Peace, here is a very useful report reviewing how China’s views on AI safety are changing, and the implications.

Data & Decision Making

An interesting article in HBR looks at how GenAI tools can be used with digital twin technology to model organizational processes and supply chains to analyse existing customer data and generate detailed virtual models of various customer segments.

Cassie Kozyrkov considers the challenges of deleting data and concludes that trying to remove training data once it has been baked into a large language model is like trying to remove sugar once it has been baked into a cake. You’ll have to trash the cake and start over.

Innovation & Collaboration

The AI product landscape might soon become even more interesting on news reported in Computer World that that former Apple designer Sir Jony Ive’s new firm LoveFrom is working with Sam Altman’s OpenAI on an AI device. It is reported that Ive has turned one of his real estate investments into the HQ of a new “artificial intelligence device company that he is developing with OpenAI.”

ChatGPT has been available for a while. But, a lot has happened in that time to that technology, and its founding company, OpenAI. To help you keep up with this, here’s a useful timeline from TechCrunch of everything you need to know about ChatGPT product updates and releases.

Productivity & Efficiency

How is AI actually being used by doctors? A survey of 1,000 GPs published in The British Medical Journal has found that 20% of them use AI tools in day-to-day clinical practice, with 29% using AI to generate documents after appointments, and 28% using AI to suggest differential diagnosis.

The latest findings from the Deloitte AI Institute’s survey series tracking GenAI adoption reviews how organizations are navigating challenges and measuring value. Two critical areas of focus are data foundations and governance, risk, and compliance.

Not sure how AI is really being used today? Here is a summary from Google of 185 real-world GenAI use cases from the world's leading organizations.

GenAI is transforming marketing and communications. Edelman’s AI landscape report looks at the quality and capabilities of the over 180 GenAI tools aimed at this market.

While we may debate the value and use of GenAI, consulting companies are pushing ahead with clients. Accenture said it had $1 billion in new GenAI bookings in its latest quarter.

Regulation and Compliance

We are still early in understanding the risks of AI. This report in The Guardian describes how the child protection agency in the state of Victoria in Australia has been ordered to ban staff from using GenAI services after a worker was found to have entered significant amounts of personal information, including the name of an at-risk child, into ChatGPT.

How can governments strike the balance between enabling innovation and protecting the public interest? Deloitte reports that outcome-based and risk-weighted regulations are an underused tool that can both protect the public interest and encourage innovation. It is sentiment echoed in this report from the Brookings Institute.

Sustainability

Is it time for a more frugal approach to AI? The AI Journal considers the challenges when we want to use AI responsibly and make sure that we use the most appropriate AI tools, considering not only the quality and precision of the outcome but also the compute cost and the associated carbon impact.

AI takes a lot of energy to run, but this podcast from Wired provides an interesting discussion on whether underwater data centres might be the answer.

Workforce & Skills

An excellent (but long!) essay from Ben Thompson looking at the way computing has changed enterprises from the mainframe to the PC to today’s first wave of AI tools. He concludes that it is the enterprise deployment of AI in the workplace that will dictate AI’s future, much like it did for computing in general.

AI adoption brings the fear of job losses. However, a blog from LSE discusses how AI can also potentially bring a big benefit to workers through enhanced workplace inclusion.